Laser Cooling of Potassium Atoms Using Collisional Redistributed Radiation

Laser cooling of a potassium-argon gas mixture using collisional redistribution of radiation Anne Saß, Ulrich Vogl, and Martin Weitz Institut für Angewandte Physik der Universität Bonn, Wegelerstraße 8, D-53115 Bonn Abstract We study laser cooling of atomic gases by collisional redistribution, a technique applicable to ultradense atomic ensembles at a pressure of a few hundred bar. Frequent collisions of an optically active atom with a buffer gas shift atoms into resonance with a far red detuned laser beam, while spontaneous decay occurs close to the unperturbed resonance frequency. In such an excitation cycle, a kinetic energy of the order of the thermal energy kBT is extracted from the sample. Here we report of recent experiments investigating the cooling of a potassium- argon gas mixture, which compared to an rubidium-argon mixture investigated in earlier experiments has a smaller fine structure of the optically active alkali atom. We observe a relative cooling of the potassium-argon gas mixture by 120 K. 1. Introduction Light as a tool for cooling of matter was considered for the first time already in 1929 by Pringsheim [1]. In the last 25 years, Doppler cooling of dilute atomic gases has developed as a very widely used application of this concept [2, 3], and more recently anti-Stokes fluorescence cooling in multilevel systems has been investigated, allowing for an optical refrigeration of solids [4-6]. In this paper, we study the cooling of an ultradense alkali - noble gas mixture by collisional redistribution of radiation. Redistribution of atomic fluorescence is most widely known in the context of magneto-optic trapping of ultracold atoms, where the mechanism is a primary cause of trap loss processes [7]. In the long investigated field of room-temperature atomic collisions, redistribution of fluorescence is a natural consequence of line broadening effects from collisional aided excitation [8]. In theoretical works, Berman and Stenholm proposed laser cooling and heating of two-level systems based on the energy loss respectively during collisional aided excitation of atoms [9]. Experiments carried out with gases of moderate density have observed a heating for blue detuned radiation, but the cooling regime was never reached [10]. In a recent experiment, our group has demonstrated the relative cooling of a gas of rubidium atoms with 230 bar of argon atoms by 66 K by collisional redistribution of radiation [11]. The cooled gas has a density of more than ten orders of magnitude above the typical values in Doppler cooling experiments. We here report on experiments where a potassium-argon mixture was used for cooling. The potassium atom has a smaller D-lines fine structure in the electronically excited level than the rubidium atom, which equals 3 nm in the case of the potassium atom instead of the value of 14 nm for the rubidium case. The pressure broadened linewidth in the high pressure buffer gas system reaches a few nm (it approaches the thermal energy kBT in wavelength units), so that there is hope that the smaller fine structure splitting of the potassium atom makes the system closer to a two-level system, which in the future can facilitate the interpretation of results. We describe the current status of the experiment and give an overview of results obtained in the potassium - argon system. In the following, Section 2 describes the principle of the cooling method, Section 3 the used experimental setup and Section 4 contains the obtained results. Finally, Section 5 gives conclusions and an outlook. 2. Cooling principle We begin by describing the principle of redistribution laser cooling of the dense gas system. Our experiment uses potassium atoms at thermal vapour pressure in 200 bar of argon buffer gas. Fig. 1 indicates the variation of the potassium 4s ground state and the 4p excited state respectively versus distance from an argon gas perturber in a binary collisional picture. Note that the potential curves are sketched schematically, a more realistic description can be found in [12]. The used cooling laser was detuned from the atomic resonance by a few nanometers to the red, i.e. 15 nm from the unperturbed potassium D1-line. Excitation becomes feasible when a noble-gas atom approaches one alkali atom and transiently shifts the D-lines into resonance. The 28 ns natural lifetime of the potassium 4p state is four magnitudes larger than the duration of a collisional process. The collisional rate in the high pressure buffer gas environment is ≈1011/s. We here assume quasi-elastic collisions, in a sense that no quenching effects occur, and a range of the potential much smaller than the interatomic spacing. Subsequent spontaneous decay occurs most likely at larger distance of atoms and perturbers, i.e. close to the unperturbed resonance frequency, see Fig. 1. During each excitation cycle, a kinetic energy of order of the thermal energy kBT is extracted from the dense atomic sample, which is of the same order of magnitude than the used laser detuning in energy units. Note that early experiments aiming towards a redistribution cooling in the moderate pressure regime were limited to detunings from resonance much below the possible energy exchange of kBT, due to the smaller value of the pressure broadening [10]. In general, the here described simple picture neglects multiparticle collisions, which can be relevant at the used buffer gas pressures of a few 100 bar used in the present experiments [14]. 3. Experimental Setup The experiment takes place in a stainless-steel high pressure chamber with sapphire optical windows. Because of its lower vapour pressure at constant temperature with respect to the rubidium atom, in the here investigated potassium case the pressure cell has to be heated to higher temperatures then in the rubidium case reported earlier [13], to achieve a sufficient optical density for the absorption of the cooling laser beam. For the cooling experiments, the 21 -3 atomic ensemble (200 bar argon and thermal vapour pressure of potassium, nAr ≈ 10 cm , nP ≈ 1016 cm-3) was initially heated up to 440°C to achieve sufficient thermal potassium vapour pressure to be optically dense. To reduce quenching of excited state potassium atoms by collisions with impurities in the buffer gas, we use argon buffer gas of purity 6.0. Cooling radiation near the potassium D-lines is generated with a continous-wave Ti:Sapphire ring laser, whose output power is 3 W. The incident laser beam is guided through the cell volume through two sapphire optical windows with 18 mm diameter. The optical path length inside the cell is 10 mm. Figure 2 shows the experimental setups for the recording of fluorescence spectra (Fig. 2a) and temperature measurements inside the cell (Fig. 2b) respectively, as described in the following. A confocal setup is used to record spectra of fluorescence of the high pressure sample. To suppress a background of scattered light backscattered mostly from the cell windows from detection, both the incident beam and the outgoing fluorescence are spatially filtered with pinholes. The fluorescence signal is subsequently detected by an optical spectrum analyzer. For a measurement of the temperature change in the gas cell induced by the cooling beam, we employed thermal deflection spectroscopy [17, 18]. For this measurement of the cooling, the Ti:sapphire laser was tuned to a wavelength of 785 nm, far to the red of the unperturbed potassium D-lines (that is 766 nm and 770 nm, for the D2- and D1- lines respectively). Inside a telescope setup, an optical chopper is used to periodically block the cooling laser beam. Thermal deflection spectroscopy of the temperature change of the gas induced by the cooling beam near its beam path is applied by directing a non-resonant helium- neon laser beam collinear to the cooling laser path, with a variable lateral spatial offset. The temperature gradient induced by the cooling beam causes a spatial variation of the refractive index, resulting in prismatic deflection of the probe beam. Its angular deflection can be detected by a position sensitive photodiode placed behind the cell. Both the cooling laser beam and atomic fluorescence are blocked by a shortpass filter located in front of the photodiode. With a translation stage, the transverse position of the cooling laser beam can be varied with respect to the probe laser beam, so that the temperature profile induced by the cooling beam in the gas can be mapped out. 4. Results and Discussion In the following, experimental results for the spectroscopy of potassium atoms in the high pressure buffer gas environment and that of temperature measurements based on thermal deflection spectroscopy measurements will be discussed. 4.1. Atomic Fluorescence Using the described confocal setup, we recorded atomic fluorescence spectra from the dense buffer gas broadened sample. The spectra were measured under constant high-pressure buffer gas conditions for a variable incident laser detuning. Figure 3 presents two typical spectra with incident wavelengths of 765 nm and 780 nm respectively. The laser beam power was 630mW on the 3µm beam radius used for this measurement. For both measurements, a clear redistribution of the wavelength to the center of the D-lines is observed. Note that the D-lines center are pressure shifted to roughly 770nm and 773nm wavelength for the D2- and D1-lines respectively, i.e. by some 3-4 nm to the red compared to the position of the unperturbed case. The spectra also show a peak at the spectral position of the incident light, which is attributed to residual scattered light from the incident laser beam. This could be further reduced with improved spatial filtering (the present experiment due to low optical density of the potassium vapour used somewhat larger pinholes than required for a true mode-matched confocal filtering, given the low intensity of the backscattered light).

Load more